Sophie Scott
Institute of Cognitive Neuroscience, UCL, London, UK

https://doi.org/10.36866/pn.103.34

Human vocal communication is primarily studied in the form of human speech – a remarkable talent and evolutionarily highly specialised motor act that involves high levels of precise motor control over the articulators and over breathing. However, we do not solely communicate vocally with speech: when we are in the grips of more extreme emotion, we frequently start to produce non-verbal vocalisations, often in a relatively involuntary fashion. This includes vocal behaviours such as screaming, sobbing and laughing.

Institute of Cognitive Neuroscience, UCL, London, UK

I first started working with these kinds of vocal acts in the 1990s, when I was collaborating with colleagues who were studying neuropsychological patients who had specific deficits in the perception of emotions. They were looking at the perception and recognition of facial expressions of emotion, and wanted to extend this to other sensory modalities: my input was to investigate their perception of emotion from the voice. I first used emotionally inflected speech but quickly moved to non-verbal emotional expressions as they contained no lexical information and were therefore better analogues for the facial stimuli. This work was successful and the same patients who had difficulties recognising a frightened or an angry face, also struggled to identify a scream as frightened or a growl as an angry expression.

Although this started as a pragmatic solution to a problem posed by a population of subjects (how to test emotions in another modality), it soon became clear that there is very little work on these kinds of non-verbal emotional expressions. The majority of emotional vocalisation research focuses on emotional speech. However, non verbal emotional expressions such as laughter and screaming are more like animals calls than they are like speech: they require very little supralaryngeal articulation, are made using primarily laryngeal and breath control mechanisms, and are produced at the same time as the facial expressions which are more commonly studied.

Acoustically, we found particular profiles of acoustic properties that were associated with the perception of different kinds of non-verbal emotional expressions – the perception of laughter, for example, was strongly associated with the highly modulated amplitude envelope that results from the characteristic ‘ha ha ha’ sound.

This in turn is driven by the involvement of the intercostal muscles: normally used smoothly to pull air into and out of the lungs during metabolic breathing, and to produce a constant sub glottal pressure, to vibrate the vocal folds during speech and song, the intercostal muscles and diaphragm start to produce large contractions during laughter, each of which contributes to a single ‘ha’ burst, as air is forcibly exhaled (NB it is also possible for these contractions to be largely acoustically silent).

If these contractions start to run into one another, then the laughter can start to sound more like silent wheezing. From this perspective, laughter is more like a different way of breathing than it is a different way of speaking. Another physiological change is a constricting of the pharynx, meaning that some sounds are made during laughter as a consequence of this constriction (e.g. glottal whistles). The intercostal contractions made during laughter are much greater than those used to control breathing during speech production, and this also affects the noises made during laughter, with very high pitched noises being produced, which would be difficult or unlikely to produce under voluntary control. My laugh can be very high pitched, and I can hit pitches when laughing that I would be unable to produce while singing.

Laughter is an interesting human behaviour to study, even in isolation: it appears to be a universal emotional expression, however claims that only humans laugh have transpired to be incorrect. Laughter has been reported in gorillas, chimpanzees and orangutans, where it can look and sound quite similar to human laughter. However, we are unable to hear many of the noises made by other animals, meaning that there may be many more examples out there: it’s also probably true that no one is out there looking for laughter. Certainly a vocal behaviour which is contextually identical to laughter has been described in rats: rats make a distinctive chirping sound when they are playing together, and when they are being tickled, and when they are anticipating being tickled. Indeed, at its heart, Panksepp has argued, laughter can be considered an invitation to play.

As all mammals play when juveniles, and some continue to play through into adulthood (dogs, humans, otters), this argument would suggest that laughter is likely to be widely found across mammals. This role in play seems counterintuitive to humans adults, who strongly associate laughter with humour, jokes and comedy, however Robert Provine has shown that even in humans, laughter is primarily a social behaviour, which is strongly primed by other people – we are 30 times more likely to laugh with someone else than if we are on our own.

What this means in practice is that we are laughing mostly when we are in the company of others – and we are still not laughing at jokes. Indeed we laugh mostly at comments and statements and although we report laughing because we are amused, we are laughing to show that we like people, understand them, agree with them, are affiliated to them as much as if not more than because something is ‘funny’. Within conversations, laughter is very tightly co-ordinated, with members of a conversation laughing together at the end of sentences, even if the conversation is in sign language rather than a vocal language, and in theory people could be laughing all the way through if they wished to. We also laugh much more often than we report: all studies that have compared actual to reported laughter find that people laugh more than they say they do. Indeed, laughter is probably the most commonly encountered non-verbal vocal emotional expression, occurring at around 7 times per 10 minutes of conversation. Provine has also noted that laughter is highly behaviourally contagious, and people will frequently laugh simply because others are laughing. Like other such contagious behaviours, such as yawning, contagious laughter is modified by social factors, and people are much more likely to catch a laugh (or a yawn) from someone they know than from a stranger.

We have found, at a cortical level, that laughter leads to a much larger amount of orofacial mirror system activation than a negative vocal expression emotion like disgust: we interpreted this as due to the social use of laughter, and there was an implication that this might relate to behavioural contagion: the perception of yawning leads to very similar patterns of orofacial mirror actuvation, and activation within this network correlates with the rated contagiousness of the yawns.

We investigated this further using two different kinds of laugh – spontaneous laughter, and more controlled, posed laughter. This was inspired by a literature showing facial differences between spontaneous and controlled smiling, and by findings that chimpanzees laugh differently if they are being tickled, or are trying to make play last longer. We recorded people laughing helplessly, and also laughing in a pleasant but not spontaneous fashion. Acoustically, these laughs are quite distinct, with the spontaneous laughter being longer and higher in pitch than the posed laughter, and the posed laughter frequently being nasalised in a way that the spontaneous laughter was not. In an fMRI study, we found significant differences in the cortical responses, with the spontaneous laughter leading to greater activation in auditory cortices, and the posed laughter leading to greater activation in medial prefrontal cortex and the anterior thalamus. We interpret this as showing a greater response in auditory fields to the genuinely novel sounds of authentic laughter, while the posed laughter was associated with more extensive processing for meaning, and was associated with activity seen in explicit mentalising tasks (although no overt response was required). This likely reflects the importance of laughter, as even when people are listening to sounds passively in in MRI scanner, they are trying to understand the laughter that they hear.

We did not find the greater activation in orofacial mirror regions that we had expected to see to the spontaneous laughter: we assumed that as it is rated as more contagious than the posed laughter, these laughs would be associated with greater activation reflecting such priming, following our earlier study. What we found instead was considerable orofacial mirror responses to both spontaneous and posed laughter, and that this was associated across participants, with their post-scan scores on a test of laughter perception. There was a positive correlation, with higher post-scan test accuracy in distinguishing spontaneous from posed laughter correlating with greater activation within peak orofacial mirror responses. This suggests that the orofacial mirror responses may not be solely linked to contagion, but that being primed to respond to any laughter might improve understanding of what the laughter means.

We do not currently know what the behavioural and affective consequences of accuracy in laughter perception might be: we are currently exploring this in studies with people who are experiencing depression, and in young adults who have conduct disorders. However, in addition to being an important social emotion, research is now showing that people in close emotional relationships use positive affective expressions, commonly laughter, to regulate their emotional states, together. Couples who manage stressful situations with laughter not only start to feel less stressed (according to physiological measures) but over a longer time scale, are the couples who are more satisfied in their relationships, and stay together for longer.

These are still relatively early days for laughter science: a search on Web of Science using the terms ‘emotion expression fear’ returns 6863 papers, while the term ‘emotion expression laughter’ returns 188. Much more needs to be done, and an important step is to see the value in laughter research for the wider community. Rather than being a necessarily trivial or silly thing to study, laughter is a very commonly encountered emotional expression, which can let us map between emotional processing and social interactions. I would argue that is time to start taking the science of laughter seriously.